We can remove above Liechtenstein, Somalia & Syria from the data set as it is not going to impact the results

CountryID, Country Name, WEBNAME, Region, World Rank, Region Rank, can be eliminated as they will not have any relevant role in forming the clusters so we remove them

we are looking to predict Freedom Score, so select 2019 Score as our y-variable and define X as the remaining values.

According to the coefficient of determination, the model explains ~99% of the variation in the data.

Minimize the number of principal components so that we can get the 95% of variance

Looking at the elbow method we can take K=6, so let's apply Kmeans with k=6

The above are the centroids for the different clusters

Optimum number of clusters should be 6.